089%4356/93 $6.00 + 0.

00

J Clio itpidemiol Vol. 46, No. 12, pp. 1417-1432, 1993

Copyright0 1993PergamonPressLtd

Printedin Great Britain.All rightsreserved

CROSS-CULTURAL
ADAPTATION OF HEALTH-RELATED
QUALITY OF LIFE MEASURES: LITERATURE REVIEW
AND PROPOSED GUIDELINES
FRANCIS GUILLEMIN,* CLAIRE BOMBARDIER*~~
and DORCAS BEATON~
School of Public Health, University of Nancy, France, *ClinicalEpidemiology Division, Wellesley
Hospital, University of Toronto and fOntario Worker Compensation Institute, Toronto, Canada
(Received in revised form

11 June

1993)

Abstract-Clinicians
and researchers without a suitable health-related quality of life
(HRQOL) measure in their own language have two choices: (1) to develop a new
measure, or (2) to modify a measure previously validated in another language, known
as a cross-cultural adaptation process. We propose a set of standardized guidelines for
this process based on previous research in psychology and sociology and on published
methodological frameworks. These guidelines include recommendations for obtaining
semantic, idiomatic, experiential and conceptual equivalence in translation by using
back-translation techniques and committee review, pre-testing techniques and re-examining the weights of scores. We applied these guidelines to 17 cross-cultural adaptation
of HRQOL measures identified through a comprehensive literature review. The
reporting standards varied across studies but agreement between raters in their ratings
of the studies was substantial to almost perfect (weighted K = 0.66-0.93) suggesting that
the guidelines are easy to apply. Further research is necessary in order to delineate
essential versus optional steps in the adaptation process.
Health
Quality of life
Validity
Guidelines

status

Cross-cultural

index

RATIONALE

body of research has recently been

devoted to the development of health-related
quality of life (HRQOL) measures. In 1991
alone, over 160 different measures were used in
the published literature [l]. Such techniques are
increasingly used in clinical trials [2,3] to determine the impact of medical intervention on
quality of life (QOL), and by public health
researchers [4] to assess the outcome of health
care services. With a few exceptions [5,6] all the
measures so far developed are in the English
language and are intended for use in English-

A large

*All correspondence should be addressed to: Dr Francis

Guillemin, Ecole de Sand Publique, Facultk de
Mbdecine BP 184, 54 505 Vandoeuvre-les-Nancy Cedex,
France.

comparison

Culture

speaking countries. There is nonetheless a need

for measures specifically designed to be used in
non English-speaking countries and also among
immigrant populations, since cultural groups
vary in disease expression and in their use of
various health care systems. This need has become more acute with the growing number of
large multicentre multicountry trials.
In order to meet that need, two options are
available: (1) to develop a new measure, and (2)
to use a measure previously developed in
another language. The first option, the generation of a new HRQOL measure is a timeconsuming process in which the bulk of the
effort is devoted to the conceptualization of the
measure and the selection and reduction of its
items. In the second option, if the transposition
of a measure from its original cultural context

1417

1418

FRANCIS
GUILLEMIN
et al.

is done by simple translation it is unlikely to be

successful because of language and cultural
differences [7]. Furthermore, the perception of
QOL and the ways in which health problems are
expressed vary from culture to culture [S]. To be
successful this option requires a systematic approach to the translation and to the cross-cultural adaptation process of HRQOL measures.
A recent effort was made by Hunt and the
European Group for Health Measurement and
Quality of Life Assessment through a crosscultural adaptation of the Nottingham Health
Profile (NHP) to several European countries
using a systematic method [9]. This effort of
standardization needs to be expanded.
In this paper, we propose a set of standardized guidelines for the cross-cultural adaptation
of HRQOL measures based on previous research in psychology and sociology [l&18] and
on published methodological frameworks for
HRQOL validity [19,20]. We review the published literature on the cross-cultural adaptation
of HRQOL instruments and evaluate the
practicality of our proposed guidelines to this
literature.

METHODS

Literature search strategy

Relevant papers reflecting the methods used
for cross-cultural adaptation were identified
from three databases: Medline (1966-1992),
Health Planning and Administration (1975-1992)
and Embase (Excerpta Medica) (1990-1992).
The search strategy used was to identify articles
with quality of life, health status, health
status indicator, functional status, questionnaires and interviews as main subject
headings (exploded) or text words in titles and
abstracts. This was matched with crosscultural, cross-cultural comparison, translation and languages as main subject
headings (exploded) or text words. Papers published between January 1966 and October 1992
were considered, without language restriction.
All references obtained were entered into Reference Manager computer software [21] to check
for duplication.
Development of the guidelines
The literature review identified several publications in the field of psychology and sociology
addressing the methodology of cross-cultural
adaptation. For example, the General Health

Questionnaire has been translated into at least

36 languages [lo] and the State Trait Anxiety
Inventory into at least 21 languages [I 11.
Although these papers were excluded from our
formal evaluation because they addressed
mainly mental health, their work suggests
methodological approaches, developed to overcome the inadequacy of simple translation,
which may be useful in the cross-cultural adaptation of HRQOL measures. We have developed guidelines and a scoring method which can
be applied in a standardized manner to evaluate
the quality of cross-cultural adaptations of
HRQOL measures. This system was based on
both empirical and theoretical findings extracted from the literature. The empirical basis
was derived from a systematic review of the
published work on cross-cultural adaptation.
Theoretical foundations
were gained from
guidelines on the methodology of assessing the
validity of HRQOL measures [19,20].
Selection of articles for review
Our literature search identified 712 references
(Table 1). Their titles and abstracts were reviewed by one author (FG) for relevance to the
study. Papers were included if they contained a
description of the methodological process used
to adapt a HRQOL measure from a source to
a target culture. Papers were excluded if they
presented only results of cross-cultural comparisons, or simply mentioned the use of HRQOL
measures in different countries or in international trials without describing the translation
and adaptation process. Papers concerning instruments to measure only pain, symptoms or
mental status/disorders as well as utilities were
also excluded.
During the selection process, any reference to
methodology quoted in these papers were used
to conduct a supplementary search in the
Science Citation Index (1980-1992) for additional material on cross-cultural adaptation of
HRQOL instruments. From 1966 to 1992, only
32 papers met the inclusion and exclusion criteria for review. Six of these publications were
in abstracts form [22-271 and were not included
in the present review since not enough information was available to assess the quality of the
adaptation process. Seven other papers [28-341
were rejected during the review process as they
were judged by subsequent raters (CB, DB) not
to have met the original selection criteria, i.e.
either not dealing with a quality of life measure
or not with the cross-cultural adaptation of such

Development of guidelines

Science Citation Index

(1988-92)

References on methodology for 4

cross-cultural adaptation in
psychology and sociology

!I\i

/Exct

criteria

L
188 refs

Excerpta Medica
(1990-92)

2 articles in non-English language

7 articles not meeting inclusion criteria

6 abstracts

) Rejection

17 articles selected for review

I
32 references

uin

L
Title and abstract review

Reference ma!,

= 712 references J

4
273 refs

5
573 refs

~c-3

Health Planning
and Administration
(1975-92)

Medlinc
(1966-92)

Table 1. Seiection of articles in the databases

1420

FRANCISGUILLIMN

measure or not containing a description of the

method used for cross-cultural adaptation. Of
the remaining 19 articles, only 2 were excluded
[35,36] by restricting our review to English
language papers. Thus, the review of 17 studies
was completed [37-531.

et al,

and using another language in the same country

(immigrants) differ, agreement between the
judges was also considered separately for each
subgroup of studies.
RESULTS

Application of the guidelines to assess quality of

studies

Settings for Cross-cultural Adaptation

The proposed guidelines include 5 sections:

(1) translations and (2) back-translations by
qualified people, (3) committee review of those
translations and back-translations,
(4) pretesting
for
equivalence
using
adequate
techniques (with bilingual or monolingual individuals), and (5) reexamination of the weighting
of scores, if relevant. If an instrument was
adapted from one culture to another using a
similar language (from American English,
U.S.A., to British English, U.K.), the steps of
translation and back-translation obviously were
not required and therefore were not assessed.
The quality of each study was assessed by two
of the investigators (DB, CB) blinded to
authors names, journal titles, and city or area
of the study. Each investigator was provided
with an operational definition of the evaluation
criteria for each section and was asked to rate
each section as good, moderate, poor
quality or not done using a standard data
extraction form (Appendix). Agreement between the two judges in these ratings was assessed using the weighted kappa statistic for
categorical judgement [54]. For each section, a
mean score across studies was calculated as the
mean of quality ratings assigned for that section
with the following values: good = 3; moderate = 2; poor = 1; if the section was rated as
not done the study was not included in this
calculation.
An overall score was also calculated for each
study by adding the scores across all sections for
that study and dividing by the number of sections. For this calculation a value of 0 was
assigned to sections rated as not done. If
translation and restoring were not considered
relevant for a particular study, these sections
were omitted from the calculation of the overall
score for that study. Agreement between judges
on these overall scores was assessed using the
intra-class correlation coefficient for continuous
data [55]. Recognizing that the concepts and
techniques involved in adapting measures for
people using similar language in another culture, using another language in another country

Our literature review identified several settings for the cross-cultural adaptation process.
A range of situations may be encountered depending on similarities and differences between
the cultures and languages of the populations
concerned. An instrument originally developed
in the English language in the U.S.A. can
readily be used by a majority of the American
population (Table 2: example 1).
Immigrants using the same language may
encounter particular problems in expressing
themselves with regard to health and HRQOL.
Therefore, particular attention should be paid
to the adaptation of cross-cultural HRQOL
measures to such populations. Immigrants to
the U.S.A., for instance Hispanics, will judge
their health and related QOL according to their
cultural origin and their degree of assimilation
into the host culture. They may have been
settled for long enough to have mastered the
English language sufficiently well to answer the
original instrument, and still refer to their
Spanish culture in assessing their situation
(Table 2: example 2).
An instrument used in a country other than
that in which it was developed may require
adaptation if the populations concerned have
another culture with similar language. For instance British English should be used in Great
Britain rather than American English. There are
sufficiently meaningful differences between the
British and American cultures to necessitate
modification of some items and validation of the
measure in its new setting [37-391 (Table 2:
example 3).
Recently settled immigrants with a low degree
of acculturation may require an instrument that
is cross-culturally adapted to their Spanish
(native) language and culture, but appropriate
to the American situation (Table 2: example 4).
Under most circumstances, instruments require adaptation for use in a different country
with both a different culture and a different
language. For instance, the American measure
would need to be modified for use in the French
language in France or in Canada (Table 2:

Cross-cultural Adaptation of Health-related QOL Measures

1421

example 5). The degree of adaptation required

depends on similarities in language structure
(there are fewer differences between most of the
European languages than there are between
European and Arabic or Asian languages) and
in culture [12].
Cross-cultural adaptation
has two components: the translation
of the HRQOL
measure and its adaptation, i.e. a combination
of the literal translation of individual words and
sentences from one language to another and an
adaptation with regard to idiom, and to cultural
context and lifestyle. Translation and adaptation are required for examples 4 and 5 while
only cross-cultural adaptation is necessary in
examples 2 and 3. The quality of the adapted
measure is then assessed with regard to its
sensibility. The elements of sensibility, as
defined by Feinstein, which need to be considered include the purpose of the measure, its
comprehensibility, its content and face validity,
its replicability and the suitability of the
scales [19].
Guidelines for Cross-cultural Adaptation

The following guidelines concentrate on the

points that must be addressed in order to preserve the sensibility of the tool in the target
culture. Table 3 summarizes the steps that are
essential in order to ensure the quality of the
procedure.
1. Translation
Produce several translations. Translations are
of higher quality when undertaken by at least
two independent translators. This allows for the
detection of errors and divergent interpretations
of ambiguous items in the original. The quality
will be even higher if each translation is undertaken by teams rather than single individuals,
who are more likely to introduce personal idiosyncrasies.
Use qualified translators. The qualifications
and characteristics of the translators are also
important. Highly educated individuals may not
be culturally representative of the target population [ 131. Translators should preferably translate into their mother tongue [40]. Some of them
should be aware of the objectives underlying the
material to be translated and the concepts involved so as to offer a more reliable restitution
of the intended measurement [56,41]. Other
translators who are unaware of these objectives
and concepts may usefully elicit unexpected
meanings from the original tool.

1422

FRANCISGUILLWN et al.
Table 3. Guidelines to preserve equivalence in cross-cultural adaptation of HRQOL measures.*
1. Translation
Produce several translations
Use qualified translators
Back-translation

Produce as many back-translations

Use appropriate back-translators

as translations

Committee review

Constitute a committee to compare source and final versions

Membership of the committee should be multidisciplinary
Use structured techniques to resolve discrepancies
Modify instructions or format, modify/reject inappropriate items, generate new items
Ensure that the translation is fully comprehensible
Verify cross-cultural equivalence of source and final versions
Pre-testing

Check for equivalence in source and final versions using a pre-test technique
Either use a probe technique
Or submit the source and 6nal versions to bilingual lay people
Immigrants: Choose the language of administration or use a dual-format measure
Weighting of scores

Consider adapting the weights of scores to the cultural context

*It is not always possible to follow all the stens described here due to the design of the measure (e.g. no
weighting &ore to be examined).

2. Back-translation
Produce as many back-translations as translations. Back-translation, translating back from
the final language into the source language, has
been shown to help improve the quality of the
final version [7,14]. Each first translations
should be back-translated independently from
each other. Misunderstandings in the first translation may be amplified in the back-translation,
and thereby revealed. Failure to adapt to the
cultural target context and ambiguity in the
source version can also be uncovered.
Use appropriate back-translators.
Backtranslation is of better quality if those who do
it are fluent in the idioms and colloquial forms
of the source language, i.e. the result of their
back-translation. Thus, they should also translate into their mother tongue. Unlike some of
the fist translators, back-translators
should
preferably not be aware of the intent and concepts underlying the material. Back-translators
without a priori knowledge of the intent of the
original instrument are free of biases and expectations and their back-translation may reveal
unexpected meanings or interpretations in the
final version.
3. Committee review
Constitute a committee to compare source and
jinal versions. A committee should be constituted in order to produce a final version of the
modified measure based on the various trans-

lations and back-translations obtained as described above. Part of that committees role
should also review the introduction and instruction to the questionnaire as well as review the
scaling of responses to each question (i.e. the
translation should maintain equivalence of steps
in Likert-type scales).
Membership in the committee should be multidisciplinary. To use the analogy of the development of a new health status measure, the
committee should consist of individuals expert
in the disease(s) explored, and in the intent of
the measure and the concepts to be explored.
Bilingual members are of particular value for
such committee [ 151.
In case of a cross-cultural adaptation for an
immigrant population, individuals representative of the target group are likely to be available.
Their input is likely to result in a measure better
adapted in terms of idioms and colloquialisms
than that which would be produced by highly
educated people [9,42]. A scale referring to the
ability to speak, write, read and understand
both languages has been developed [16] and can
be useful in selecting these bilingual committee
members.
Use structured techniques to resolve discrepancies. The committee may resolve problems by
considering the material it has now collected. It
may further decide to repeat the translation-back-translation
process. A decentring
technique [17] has also been proposed as a way

Cross-cultural Adaptation of Health-related QOL Measures

@semantic equivalence

1423

is equivalence in the
meaning of words, and achieving it may present problems with vocabulary and grammar.
For example, vocabulary problems may be
encountered in the question are you able to
bend ? which can refer to several parts of the
body, such as the arm, back or knees, and
might have been intended-and
translatedonly to explore the ability to flex (arm), bend
over (back) or squat (knees). Furthermore,
some words, such as happy, have several
subtly different meanings depending on the
context.
Grammatical
alterations are sometimes
necessary in the construction of sentences. For
example, languages without the gerund form
may be more difficult to adapt [13]: activities
couched in terms such as dancing, singing or
Modifv instructions orformat, modifv or reject
eating
(gerund form of to dance, sing and eat)
inappropriate items, generate new items. The
may
not
be readily translatable.
committee must ensure that the introduction to
??idiomatic equivalence. Since idioms and collothe research tool and the instructions for filling
quialisms are rarely translatable, equivalent
in the questionnaire are carefully translated in
expressions have to be found or items have to
order to preserve the replicability of the measure
be substituted. This is more likely to be necess[19]. The redundancy principle, i.e. repeating the
ary in the emotional and social dimensions.
same instruction in a different manner, may help
For example, Do you feel downhearted and
to reduce comprehension errors [12].
blue? or Do you feel at home? are untransThe review committee is also likely to modify
latable idioms for which equivalents must be
or eliminate irrelevant, inadequate and ambigufound. The item I am feeling on edge in the
ous items and may generate substitutes better
NHP was translated into I have my nerves
fitting the cultural target situation while mainoutside my skin in Italian [9], I feel nervous,
taining the general concept of the deleted items.
tense in French [43] and I am afraid in
Ensure that the translation is fully comprehenArabic [60].
sible. Guidelines about how to produce trans??experiential equivalence. The situations evoked
lations comprehensible to a majority of people
or depicted in the source version should fit the
have suggested using language which can be
target cultural context. This may result in the
understood by 10 to 1Zyear old children [12].
modification of an item. For example, in the
Recommendations include: short sentences with
Brazilian version of the HAQ, using public
key words in each item as simple as possible [ 111;
transportations was substituted for using a
the active rather than the passive voice; repeated
private car, since most of the people in Brazil
nouns instead of pronouns; and specific rather
have no car [44]. I have forgotten what it is
than general terms. Authors should avoid using:
like to enjoy myself [60} and How many
metaphors and colloquialisms; the subjective
hours a week do you have leisure activities?
mode; adverbs and prepositions telling where
do not refer to usual experiences in a number
and when, possessive forms; words indicating
of cultures and equivalent feelings (enjoy) or
vagueness; and sentences containing two differactivities (leisure) must be found or the items
ent verbs that suggest different actions.
discarded.
Verifv cross-cultural equivalence of source and ??conceptual equivalence refers to the validity of
final versions. Several taxonomies of cross-culthe concept explored and the events experitural equivalence have been proposed in the
enced by people in the target culture, since
psychiatry literature [12, 13,60,61]. The ultiitems might be equivalent in semantic meaning
mate parity is equivalence of HRQOL concepts
but not conceptually equivalent.
within the cultures concerned. Translators aimFor example, cousin and brother may
ing for conceptual equivalence should consider
mean more than simply second or first-degree
the following:
relative of the same generation. In many culof improving cross-cultural adaptation. This
technique considers the source and final versions equally important. Both are open to
modification during the translation procedure.
In other words, the measure is not considered to
centre on one of the languages. Decentring is
best conducted in close collaboration with the
authors. If problem items are found, the authors
may provide a working version of the instrument or items, maintaining the concept of the
questions, but avoiding colloquialisms. Searching for a common way to express a concept in
both languages is the best way to ensure that the
final version maintains content validity. It is
unusual for authors to be available, and this
process may need to be conducted by committee
members.

1424

FRANCIS
GUILLEMIN
et al.

tures in developing countries they have a wider

meaning within the social network. I have
pain in my head may translate perfectly into
another language semantically, but have a
totally different conceptual meaning for the
target culture [60].
4. Pre -testing
Check for equivalence in source and final
versions using a pre-test technique. In pretesting, a sample population replies to the
questionnaire in order to check for errors and
deviations in the translation. Two techniques
are available: a probe technique and appraisal
by bilingual individuals. Both allow for the
checking of face validity, i.e. the confirmation
that questions are acceptable without arousing
reluctance or hesitation. If the final version
does not achieve a satisfactory level of equivalence, further revision can be performed by the
committee.
Either use a probe technique. The answer to
an item might appear adequate, yet be consistently misunderstood. In order to determine
whether a questionnaire is being understood
correctly, it can be administered to a group of
patients as follows. After each answer (or a
random sample of answers), the patient is
asked the probe question: What do you
mean ? and is encouraged to elucidate his or
her understanding of the item in an open-ended
manner [18]. This ensures that the final item is
understood as having a meaning equivalent to
that of the source item.
Or submit the source and final versions to
bilingual lay people. The source and the final
versions of a measure can be administered to a
group of bilingual individuals in order to detect
possible discrepancies. This method can also
help pinpoint any inadequacy of the final
version with the cultural context. They are
asked to rate the equivalence of each item
between the source and final versions. Those
items with low level of equivalence or rated
discrepantly by different people can still be
revised at this stage [l 1, 121. Administration of
a questionnaire to bilingual lay people is not
practical in every setting but may be possible
with immigrants.
The case of adaptation for immigrants requires two additional considerations:
Choose the language of measurement administration for immigrants. During questionnaire
administration, some immigrant respondents
express a preference for their native language.

Several methods have been described for the

choice of language of administration; it can be
the decision of the respondent her/himself, or
of the interviewer or research assistant, or it
can be based on a measure of acculturation.
Several such measures have been proposed
[57-591 and include language proficiency and
preference, country of birth and origin, location of education, ethnic identification, contact with homeland, and ethnicity of childrens
friends, combined to form an acculturation
score. Even if not used to determine the interview language, the score can be a useful covariable when investigating several cultural groups.
Use a dual-format measure for immigrants.
On the other hand, immigrants may switch
language during an interview and responses
should be recorded in the language used (on a
two-language form). With regard to selfadministered questionnaires, the best option is
to present the material in a dual-language format, either on two separate pages [62] or item
by item [40]. The measure of HRQOL by
proxies is not recommended, even when respondents are illiterate, since individual subjective appreciation is not reliably assessed by
other raters [63]. In this situation, interview
must be preferred to self-administered questionnaire. Interview may also be appropriate
when questionnaires are not appropriate to the
respondents culture [9].

5. Weighting scores
Consider adapting the weights of scores to the
cultural context. A scoring method using
weights is provided with the source versions of
some instruments (Sickness Impact Profile [64],
NHP [65]) in order to combine the information
in an index or in several indices (profile). However, the weighting may not apply to the new
cultural situation. It can be reexamined either
by judgement or using a mathematical approach. Using judgement, the cross-cultural
validity of the weighting of items is reexamined
by experts, who may be health care professionals, patients or lay people. Several techniques are available to elicit culture-adapted
weights from expert opinion. With a mathematical approach, data obtained from a
sample of patients are analysed by various
statistical techniques for scalability (Gutmann
analysis) or dimensionality (factor analysis) in
order to work out the best way of aggregating
the information in one index or several indices.

Cross-cultural Adaptation of Health-related QOL Measures

Application of the Guidelines

Study characteristics
A description of the content of the 17 studies
analysed is given in Table 4. Some instruments
have been adapted to another language by
different research teams addressing different aspects of the adaptation. For instance, a Chicano
version of the Sickness Impact Protile (SIP,
originated in English [64]) produced by one
team [50] has been further examined for score
weights by another team [42].
The methods for adaptation reported in the
17 studies were heterogenous. In the adaptation
of the instruments from one culture in another
with the same language (three studies), i.e. from
American to British English, rewording of items
was never carried out by a specified committee.
Pretesting of the measure was conducted in two
studies using either a probe technique [37] or a
comparison of the original and final versions
[38]. The weight of scores were reassessed in one
study [39]. In the adaptation of instruments
from one culture to another which uses a different language (8 studies), the translation techniques used varied from one literal translation
to three translations performed independently
by one to two translators with varying degrees
of qualification. Back-translation was described
in six studies. Only one study with multiple
translations specified using as many back-translations as translations [41]. When a committee
for review was constituted (5 studies), its composition varied from two authors to 12 people
including physicians, other health professionals,
patients, non-patients and bilingual people.
Pretesting of the measure was conducted in four
studies. The weighting of scores, examined in
two of three studies where it was applicable
[43,471 involved the use of large samples
of patients and non-patients
as experts.
Weights were derived by expert elicitation
using the Thurstone method, as in the original
instrument.
Six papers addressed important cultural
groups of immigrants in the host country:
Asians in the U.K. and Hispanics in the U.S.A.
The translation techniques used varied considerably, from one to five translations, each involving one to three translators in one to four
back-translations. A review committee of 2 to
15 carefully selected bilingual people was constituted in five studies. A pre-test of the new
versions was conducted in three studies by 12 to
3 1 bilingual individuals. Weighting of score was

1425

reassessed in one of three relevant studies (SIP

in Spanish [42]) using expert evaluation by a
group of health care consumers, as in the original instrument.
Quality of the studies
For each of the individual guidelines, the
mean quality scores on a scale of 1 = poor,
2 = moderate and 3 = good, ranged between 1.9
and 2.4 (Table 5). Thus, when a guideline was
addressed in a study, raters judged the methods
to be of moderate quality. According to Landis
[66], there was substantial to almost perfect
agreement between the two judges in these
ratings (weighted K 0.660.93).
Since not all of the relevant guidelines were
considered in each study, the overall scores were
much lower. In calculating the overall score
when a guideline rated as not done is assigned
a value of 0, the mean quality score overall
studies was 1.3. It was higher in immigrant
populations (1.6) than in adaptation to another
country using the same or another language (0.8
and 1.3 respectively). The agreement between
judges for the overall quality score for the 17
studies was high (ICC = 0.92). It was higher in
studies where the target was another country
using the same language (ICC = 1) or adaptation to another language in another country
(ICC = 0.96) than in studies for adaptation to
immigrant populations (ICC = 0.87).
COMMENTS

Cross-cultural adaptation must be clearly distinguished from cross-cultural comparison since

the two processes rely on different research
hypotheses [67]. Adaptation is oriented towards
measuring a similar phenomenon in different
cultures; it is essentially the production of an
equivalent instrument adapted to another culture. Cross-cultural comparison refers to the
comparative study of a phenomenon across
cultures in order to identify differences attributable to culture. It is possible only after the
measurement tool has been adapted and is
equivalent in both cultures. Thus, the crosscultural adaptation of a measure is a prerequisite for the investigation of cross-cultural
differences.
The articles reviewed were mainly describing
the cross-cultural
adaptation
of Englishlanguage instruments into European languages
in European countries. Although our search
strategy included all languages, 17 of the 19

Language

Sweden

HAQ Swedish, 1988 [46]

HAQ Portugese, 1990 [44]

HAQ French, 1992 [41]

HAQ, 1980 [71]

France

Brazil

Canada

Adaptation in another language, other country

AIMS French, 1990 [45]

U.K.

SIP British, 1985 [39]

SIP, 1981 [64]

3 translations
(2 translators in
each team, aware
of objectives and
disease)

1 translation by 2
translators aware
of objectives and
intent

1 translation
1 translator aware
of objective and
disease
-

N/A

N/A

U.K.

HAQ British, 1986 [38]

N/A*

Translation

U.K.

Country

HAQ, 1980 (711

AIMS, 1980 [70]

Fru~crs GUILLEMINer al.

3 back-translations
(1 English-cultured)

1 back-translation
by 2 translators

N/A

N/A

N/A

Back-translation

2 Rheumatologists
2 English professors
2 Physiotherapist
1 Occupational
therapist
5 RA patients
Authors + 2 bilinguals

2 bilingual
researchers

Committee

64 RA patients
Check relevance
31 RA patients
with translation
22 RA patients
with substitute version
Probe technique

30 RA patients
Probe technique:
questionnaire
+ interview
27 RA patients
Check understandability of the original: 22% pbs
33 RA patients
Check understability of modified
HAQ: 6% pbs
-

Pre-testing

Table 4. Reviewed articles with description of the methodology used for the cross-cultural adaptation of HRQGL measures

Aaizptation in similar kmguage, other country

AIMS British, 1990 [371
AIMS, 1980 [70]

Griginal
instrument

1426

N/A

WA

N/A

135 consumers
(stratified by age/sex)
54 health profi 46
nurses + 8 physicians
Thurstone method

N/A

N/A

Weighting score

France
Spain

NHP French, 1990 [43]

NPH Spanish + Catalan,

1990 [49]

AIMS Spanish, 1989 [Sl]

SIP Spanish, 1980 [SO]

SIP Spanish, 1984 [52]

SIP Spanish, 1992 [42]

NHP Urdu, 1989 [53]

SIP, 1981 [64]

-

NHP, 1980 [as]

*N/A, not applicable.

AIMS Spanish, 1989 [40]

AIMS, 1980 [70]

U.K.

U.S.A.

U.S.A.
U.S.A.

U.S.A.

U.S.A.

Adaptation in another language, same country (immigrants)

Sweden

NHP Swedish, 1987 [471

NHP, 1980 [65]

Sweden

QOLS Swedish, 1992 (481

QOLS, 1978 [72]

2 tranlations by 3
translators each: 1
literal + 1 idiomatic
2 translations
(1 translator each)

several translations
4 translations: 3 by
1 translator and 1
by 2 translators

2 translations: 1
idiomatic + 1 literal (3 translators
each)
3 previous translations
1 translation (1)

2 reviewers for 2
previous + 2 new
translations
3 authors

1 back-translation
(1 translator)

bilinguals

3 English and 3

15 selected Hispanic American

Selection of best
version
Iteration of backtranslation
2 translators?

1 back-translation
by other bilinguals?

1 backtranslation (1)
4 back-translations, each by 4
bilingual translators

(1)
2 previous backtranslation

2 back-translations

29 health care consumers valuing statemerits

N/A

N/A

N/A

31 bilingual comparison to SIP

English overall r
- 0.95
(0.65-0.95)
12 bilinguals
choose the best
version

12 selected bilinguals
Check comprehension and
readability

Cardiology
+ rheumatology
patients
Check acceptability

2 experts

1 back-translation
undetailed

2 translations
(1 translator) by
experts in NHP
10 unskilled
workers

Bilinguals

1 back-translation

259 patients + relatives

Thurstone method
355 Hosp patients
270 non patients
Thurstone method
-

1 translation

N/A

In a pilot study

1 back-translation
by 1 translator

1 translation by 2
translators aware
of objective and
disease (authors)
-

ii

B
=

c1
i:
I

1428

FR+NC~SGUILLEMINet al.

Table 5. Mean scores and agreement between two

iudaesin assessmentof aualitv of cross-culturaladantationstudies

Number of
studies
Individual guidelines (5 sections)
Translation
Back-translation
committee
Pre-testing
Weighting scores
Overall guidelines
Similar language, other country
Other language, other country
Other language, same country
(immigrants)
All articles

12
10
10
9
6

Mean score* [range]

Intra-class
correlation
coefficient

2.1 [l-3]
1.9 [l-3]
2.2 [l-3]
2.1 [l-3]
2.4 [2-31

Weighted kappa
0.70
0.66
0.86
0.88
0.93

3
8
6

0.8 [0.5-l]
1.3 [O&2.1]
1.6 [OA-2.61

1
0.96
0.87

17

1.3 [O/l-2.6]

0.92

*Mean score across studies calculated as the mean quality ratings assigned for each section with the following values:
Individual guidelines: good = 3, moderate = 2, poor = 1;
Overall guidelines: good = 3, moderate = 2, poor = 1, 0 = not done.

eligible articles were papers published in the

English language. It may be that the use of
self-administered instruments to measure QOL
is an English cultural phenomena. Alternatively,
non-English papers may have been missed since
papers published only in national journals may
not be included in the three medical databases
we searched. These papers would have been
overlooked unless they were cited in another
article included in a database [35]. Finally,
much of this research may not be published
because it is not the main topic of the research
but only a preliminary step toward an application of a QOL measure. However, the increasing number of publications appearing in recent
years reflects the growing importance and interest attached to the methodology of cross-cultural adaptation.
Our review of the literature indicates a lack of
standardized approach to the cross-cultural adaptation of HRQOL instruments. The methodologies vary and often the authors do not give
the readers essential information to understand
the strength of the translation. Interestingly,
citation searches (Science Citation Index) using
methodology papers found in the psychology
and sociology literature for adaptation of tools
in these fields failed to produce additional references relating to HRQOL tools. This suggests
that many researchers in QOL may not be aware
or do not quote this methodological work developed in the psychology and sociology literature.
Based on our review of the methods of crosscultural adaptation in the field of psychology
and sociology, and our review of the HRQOL
measures, we propose a set of guidelines which
includes 5 essential steps for the translation and
cross-cultural adaptation of HRQOL measures.

The agreement between the judges using the

proposed guidelines appeared substantial to
almost perfect. The results were consistent
within and across the sections of these guidelines and in different cross-cultural adaptation
settings. This indicates that they are appropriate, easy to interpret and suggests that they can
be used further with satisfactory reliability.
The quality of the methodology employed for
the adaptation of HRQOL measures was rated
between 1.9 and 2.4 (on a l-3 scale) in each
section. However, in each study assessed, some
aspect of the adaptation process is likely to have
been underreported (even if it was carried out)
perhaps because it was not at the time regarded
as important. Hence, a score of 0 (not done)
might have underrated a work that authors
merely did not made explicit. For this reason we
are not reporting on individual paper scores.
For instance, the replicability of the adapted
instrument, i.e. the clarity of the presentation
and the thoroughness of the directions provided
for its use [19] was addressed in none of the
papers, although it might well have been
considered by authors.
Whether reliability, validity and sensitivity to
change should also be considered in the crosscultural adaptation process is a matter of controversy. Overall the 17 studies, the reliability of
the final version was assessed in 6 studies, the
construct validity in 9 studies and the responsiveness in 1 study. On the one hand, one may
think that the full achievement of cross-cultural
equivalence conveys the equivalence of the original measurement properties. But on the other
hand, one may argue that because of the adaptation process, the modified instrument has
unknown reliability, validity or sensitivity to

Cross-cultural Adaptation of Health-related QQL Measures

change in the new culture. Since this question is

not clear, we did not address this methodology.
Overall, the guidelines used here cannot yet
be taken as firm recommendations. Very little
research has been done in this field to delineate
what is essential from what is supplementary in
the process of cross-cultural adaptation. It is
important to stress that our guidelines address
only the quality of the process of adaptation.
The quality of the final product (the adapted
version) can only be judged by a qualified
committee. A single investigator or clinician can
only judge the process used in the translation
and cross-cultural adaptation since a single individual cannot fultil the linguistic and other
committee qualifications. Our literature review
suggests that each step adds quality to the final
version in terms of equivalence of concepts
explored between source and final instruments,
but this benefit has to be weighed against the
feasibility of the process.
Important factors in determining the feasibility of a particular technique are the constraints of time and resources. Bilingual people
are often in short supply. Variations in how
bilingualism is defined may mean they are even
more rare*. This problem can be more difficult
to deal with when an original instrument is
developed in a language other than English. In
some such situations, for example in international studies involving several countries in
Europe, investigators may opt to translate the
instrument into English first, as a common
communication language, and then adapt it into
the other language as required. If a shortage of
bilingual people impedes the pre-testing with
bilingual respondents, a probe technique with
monolingual respondents can be used.
The preservation of the sensibility of an instrument is time-consuming but unavoidable.
The only step that can possibly be shortened is
the examination of the weighting of scores, if
any, by simply accepting the weights of the
original instrument. It should be borne in mind,
however, that the validity of the final score may
be diminished [43].
Further research is required to establish a
*Some authors have considered as bilingual people (with
one language) who have lived at least one year in another
country (with another language) [42]. Other define
people as bilingual only when they have been reared in
two cultural and language contexts, keeping in touch
with both [40]. Although the former definition probably
does not allow for an understanding and mastery of
idioms and colloquialisms, the latter situation may well
be too rare to be useful.

1429

method to quantify the equivalence of source

and final instruments across cultures, and to
identify essential versus optional steps in the
adaptation process. Research is also needed to
determine adaptation
needs for HRQOL
measures where items are selected by the patient
(Patient Elicitation Technique [68], SEIQoL
[69]) rather that the questionnaires addressed in
this paper.
In conclusion, the adaptation of a preexisting
measure to the cultural context of a target
population, as described above, has several
advantages:
??it

provides a common measure for the investigation of HRQOL within different cultural
contexts;
??it offers a standard measure for use in international studies, many of which are now being
conducted;
??it
allows comparisons between national/
cultural groups relying on a standard measure
designed and adapted to measure the phenomenon cross-culturally;
??it allows the inclusion of immigrants avoiding
the frequent bias of representing only the
dominant culture of the country;
??it is less costly and time-consuming than generating a new measure. Nevertheless, it should
be borne in mind that the cross-cultural adaptation of HRQOL also requires careful attention, involves numerous people and is
time-consuming.
Acknowledgements-The authors are grateful to Dr Stephen
Stansfeld and Dr James Wright for fruitful comments on
earlier drafts of this manuscript. F. Guillemin was supported by a Bourse Lavoisier from the French Ministry of
Foreign Affairs and by the Fondation pour la Recherche
Therapeutique.

REFERENCES
1. Spilker B, Simpson RL, Tilson HH. Quality of life
bibliography and indexes: 1991 update. J Clin Res
WarntaeoePidemlol 1992; 6: 205-266.
2. Aaronson NK, Meverowitx BE. Bard M et al. Oualitv
of life research in~oncology. Past achievements and
future priorities. Caneer 1991; 67: 839-843.
3. Fitzpatrick R, Fletcher A, Gore S, Jones D, Spiegelhalter D, Cox D. Quality of life measures in health
care. I: Applications and issues in assessment. Br Med
J 1992; 305: 10741077.
4. Advances in Health Status Assessment: Proceedings of
a conference. Med Care 1992; 30: MSl-MS293.
5. Gerin P, Daxord A, Cialdella P, Leizorovicz A, Boissel
JP. Le questionnaire Protll de la qualite de la vie
subjective. Premiers elements de validation. Tberapie
1991; 46: 131-138.
6. Kempen GI, Suurmeijer TP. The development of a
hierarchical polychotomous ADLIADL
scale for

1430

7.

8.

9.

10.

11.

12.

13.

14.
15.

16.

17.

18.

19.

20.

21.
22.

23.

24.

25.
26.

FRANCIS
GUILLEMIN
et of.

noninstitutionalized elders. Gerontologist 1990; 30:

497-502.
Berkanovic E. The effect of inadequate language translation on Hispanics responses to health surveys. Am J
PubIic Healtk 1980; 70: 1273-276.
Kleinman A, Eisenberg L, Good B. Culture, illness
and care: clinical lessons from anthropologic and
cross-cultural research. Ann Intern Med 1978; 88:
251-258.
Hunt SM, Alonso J, Bucquet D, Niero M, Wiklund I,
McKenna S. Cross-cultural adaptation of health
measures. Health Policy 1991; 19: 3344.
Goldberg D, Williams P. A Users Guide to the General
Health QuestIomtaire. Windsor: NFER-NELSON,
1990.
Spielberger CD, Sharma S. Cross-cultural measurement of anxiety. In: Spielberger CD, Diaz-Guerrero R,
Eds. Cmm-ctdtural Anxiety. Washington: Hemisphere
Publ. Corp; 1976: 13-25.
Brislin RW, Lonner WJ, Thorndike RM. Questionnaire wording and translation. In: Cross-cuItural
Reaeareh Methods. New York: John Wiley; 1973:
32-58.
Secherst L, Fay TL, Zaidi SMH. Problems of translation in cross-cultural research. J Cross-caItural
Psycho1 1972; 3: 41-56.
Brislin RW. Back-translation for cross-cultural research. J Cross-caItaraI Psycho1 1970; 1: 185-216.
Deutscher I. Asking questions cross-culturally: some
problems of linguistic comparability. In: Warwick DP,
Oskerson S, Eds. Comparative research methods. New
Jersey: Prentice Hall; 1973: 163-186.
Gonzalez-Reigosa F. The anxiety-arousing effect of
taboo words in bilinguals. In: Spielberger CD, DiazGuerrero R, Eds. Cross+ulturaI Anxiety. Washington:
Hemisphere Pub1 Corp; 1976: 89-105.
Werner 0, Campbell DT. Translating, working through
interpreters, and the problem of decentering. In:
Narroll R, Cohen R, Eds. A Handbook of Metbad in
Cultural Anthropology. New York: American Museum
of Natural History Press; 1970: 398-420.
Schuman H. The random probe: A technique for
evaluating the quality of closed questions. Am Soeiol
Rev 1966; 31: 218-222.
Feinstein AR. The theory and evaluation of sensibility.
In: Feinstein AR, Ed. Clinimetrics. New Haven: Yale
University Press; 1987: 141-166.
Bombardier C, Tugwell PX. Methodological consideration in functional assessment. J Rheamatol 1987; 14
(Suppl 15): 610.
Reference Manager, version4. Carlsbad, CA: Research
Information System, Inc.; 1989.
Perez ER, Mackenzie CR, Ryan C. Development of a
Spanish version of the Modified Health Assessment
Questionnaire. Arthritis Rheum 1990; 33 (Suppl 9):
SIOO (Abstract).
Cardiel MH, Abello-Ban8 M, Ruiz-Mercado R,
Alarcon-Segovia D. Quality of life in rheumatoid
arthritis: validation of a Spanish version of the disability index of the Health Assessment Questionnaire.
Arthritis Rheum 1991; 34 (Suppl. 9): S183 (Abstract).
Raynauld JP, Singh G, Shhoky JB, Haraoui B, Choauette D. Blier C. Fortin P, Pelletier JP, Gascon J,
Goulet JR, Nevillk C, Fries. JF. A French-Canadian
version of the Health Assessment Questionnaire.
ArthrItIs Rheum 1992; 35 (Suppl. 9): S125 (Abstract).
Archenholtz B, Bjelle A. Reliability and validity of a
Swedish version of the Arthritis Impact Measurement
Scales. ScantI J Rkeumatol 1990: 20: 219 (Abstract).
Abello-Banfi M, Cardiel MH; Ruiz-Mercado R,
Alarcon-Segovia D. Quality of life in rheumatoid
arthritis: validation of a Spanish version of the

27.

28.
29.

30.

31.

32.

33.

34.

35.

36.

37.

38.
39.
40.

41.

42.

43.

44.

Arthritis Impact Measurement Scales. Arthritis Rheam

1991; 34 (Suppl. 9): S183 (Abstract).
Esteva M, Gonzalez N, Ruiz M. Reliability and validity of a Spanish version of the Sickness Impact
Protile. Arthritis Rheas 1992; 35 (Suppl. 9): S219
(Abstract).
Chang RI-I, Dodder RA. The modified purpose in life
scale: a cross national validity study. Int J Aging Hum
Devel 1983-84; 18: 207-217.
Kaplan RM, Bush JW, Berry CC. The reliability,
stability and generalizabihty of a health status index.
Pme So&d Statistics Section. American Statistical
Association; 1978: 704709.
Brazier JE, Harper R, Jones NMB, OCathain A,
Thomas KJ, Usherwood T, Westlake L. Validating the
SF-36 health survey questionnaire: new outcome
measure for primary care. Br Med J 1992; 305:
160-164.
Siegert CEH, Vleming LJ, Vandenbroucke JP, Cats A.
Measurement of disability in Dutch rheumatoid
arthritis patients. CBn Rkemnatol 1984; 3: 305-309.
Van der Heidje DMFM, Van Riel PLCM, Van de
Putte LBA. Sensitivity of a Dutch Health Assessment
Questionnaire
in a trial comparing
hydroxychloroquine vs sulphasalazine. ScareI J Rheamatol
1990; 19: 407412.
Cavalieri F, Salal F, Ferracioli GF. Relationship
between physical impairment, psychosocial variables
and pain in rheumatoid disability. An analysis of their
relative impact. Clin Exp Rheamatol 1991; 9: 47-50.
Wiklund I, Romanus B, Hunt SM. Self-assessed disability in patients with arthrosis of the hip joint.
Reliability of the Swedish version of the Nottingham
Health Profile. Int Disabil Studies 1988; 10: 1599163.
Ferracioli GF, Cavalieri F, Salaffi F, Mercadanti M.
Compromissione funzionale, invalidita ed handicap
nellartrite reumatoide. Studio prospettico mediante
questionari orientari. Reumatismo 1988; 40: 45-56.
Jacobs HM, Luttik A, Touw-Otten FW, de Melker
RA. The Sickness Impact Profile (SIP); results of an
evaluation study of the Dutch version. Nederlands
Tijdscbift Vaor Geneeskunde 1990; 134: 1950-1954.
Hill J, Bird HA, Lawton CW, Wright V. The Arthritis
Impact Measurement Scales: An anglicized version to
assess the outcome of British patients with rheumatoid
arthritis. Br J Rheumatol 1990; 29: 193-196.
Kirwan JR, Reeback JS. Stanford Health Assessment
Questionnaire modified to assess disability in British
patients with RA. Br J Rkeumatol 1986; 25: 206-209.
Patrick DL. Sittamoalam Y. Sommerville SM. Carter
WB. Berg& M. A cross-cultural comparison of health
status values. Ant J Public Healtk 1985; 75: 1402-1408.
Hendricson WD, Russel IJ, Jacobson JM, Rogan H,
Bishop GD, Castillo R. Development and initial validation of a dual language English-Spanish format for
the Arthritis Impact Measurement Scales. Arthritis
Rheam 1989; 32:-1153-1159.
Guillemin F, Brianqon S, Pourel J. Validity and discriminant ability of a French version of the Health
Assessment Questionnaire in early RA. Disabii Rehabil
1992; 14: 71-77.
Deyo RA. Pitfalls in measuring the health status of
Mexican Americans: comparative validity of the Enelish and Soanish Sickness Imoact Profile. Am J Public
HeaIth 198b; 74: 569-573. Bucquet D, Condon S, Ritchie K. The French version
of the Nottingham Health Profile. A comparison of
items weights with those of the source version. See Sci
Med 1990; 30: 829835.
Bosi-Ferraz M, Magalhaes Oliveira L, Araujo PMP,
Atra E, Tugwell P. Cross-cultural reliability of the
physical ability dimension of the Health Assessment
Questionnaire. J Rbeumatol 1990, 17: 813-817.

Cross-cultural Adaptation of Health-related QGL Measures

45.

46.
41.

48.

49.
50.

51.
52.

53.
54.
55.
56.
51.

58.

Sampalis JS, Pouchot J, Beaudet F, Carette S,

Gutkowski A. Harth M. Mvhal D. Seneca1 JL, Yeadon
C, Williams Ji, Esdaile jM: Arthritis Impact &asurement Scales: reliability of a French version and validity
in adult Stills disease. J Rheumatol 1990; 17:
1657-1661.
Ekdahl C, Eberhardt K, Andersson ST, Svensson B.
Assessing disability in patients with rheumatoid
arthritis.-Seand J dheumaiol 1988; 17: 263-271.
Hunt SM. Wiklund I. Cross-cultural variation in the
weighting of health statements: a comparison of
English and Swedish valuations. Health Policy 1987; 8:
227-235.
Burckhardt CS, Archenholtz B, Bjelle A. Measuring
the quality of life of women with rheumatoid arthritis
or systemic lupus erythematosus: a Swedish version of
the Quality of Life Scale (QOLS). Stand J Rheumatol
1992; 21: 190-195.
Alonso J, Anto JM, Moreno C. Spanish version of the
Nottingham Health Profile: translation and preliminary validity. Am J Public Health 1990; 80: 704-708.
Gilson BS, Erickson D, Chavez CT, Bobbitt RA,
Bergner M, Carter W. A Chicano version of the
Sickness Imuact Profile. Cult Med Psvchlat 1980: 4:
137-150. Coulton CJ, Hyduk CM, Chow JC. An assessment of
the Arthritis Impact Measurement Scales in 3 ethnic
groups. J Rheumatol 1989; 16: lll(t1115.
Hendricson WD, Russel IJ, Prihoda TJ, Jacobson JM,
Rogan A, Bishop GD. An approach to developing a
valid Spanish language translation of a health status
questionnaire. Med Care 1989; 27: 959-966.
Ahmad WIU, Kemohan EEM, Baker MR. Cross-cultural use of socio-medical indicators with British
Asians. Health Policy 1989; 13: 95-102.
Fleiss JL. Ihe Design and Analysis of Clinical Experiments. New York: Wiley; 1986.
Dunn G. Desin and Analysis of Rellabillty Studies. New
York: Oxford Universitv Press: 1989.
Del Greco L, Walop G, Eastiidge L. Questionnaire
development. 3. Translation. Can Med Assoc J 1987;
136: 817-818.
Chesney AP, Chavira JA, Hall RP, Gary HE. Barriers
to medical care of Mexican-Americans: the role of
social class, acculturation and social isolation. Med
Care 1982; 20: 883-891.
Schribner R. Dwver JH. Acculturation and low birthweight amodg Laiinos in the Hispanic HANES. Am J
Public Health 1989; 79: 1263-1267.

1431

59. Marks G, Solis J, Richardson JL, Collins LM, Birba

L, Hisserich JC. Health behavior of elderly Hispanic
women: does cultural assimilation make a difference?
Am J Public Health 1987; 7: 1315-1319.
60. Hunt SM. Cross-cultural issues in the use of sociomedical indicators. Health Policy 1986; 6: 149-158.
61. Flaherty JA, Gaviria FM, Pathak D, Mitchell T,
Wintrob R, Richman JA, Birz S. Developing instruments for cross-cultural psychiatric research. J New
Ment LX.91988; 176: 257-263.
62. Karno M, Bumam A, Escobar JI, Hough RL, Eaton
WW. Development of the Spanish-language version
of the national institute of mental health diagnostic
interview schedule. Ard~ Gen Psycbiat 1983; 40:
1183-l 188.
63. Sprangers MA, Aaronson NK. The role of health care
providers and signiiicant others in evaluating the
quality of life of patients with chronic disease: a review.
J CIin Epidemiol 1992; 45: 743-760.
64. Bergner M, Bobbitt RA, Carter WB, Gilson BS. The
Sickness Impact Profile: Development and final revision of a health status measure. Med Care 1981; 19:
787-805.
65. Hunt SM, McKenna SP, McEwen J. A quantitative
approach to perceived health status: a validation
&dy. J Epide&ol Commun Health 1980; 34: 281-286.
66. Landis JR. Koch GG. The measurement of observer
agreementfor categorical data. Biometrlea 1977; 33:
159-174.
67. Campos SS, Johnson TM. Cultural considerations. In:
Spilker B, Ed. QuaIlty of Life IO CIinical Trials. New
York: Raven Press; 1990: 163-170.
68. Tugwell PX, Bombardier C, Gent M, Bennett KJ,
Bensen WG, Carette S, Chalmers A, Esdaile JM,
Klinkhoff AV, Kraag GR, Ludwin D, Roberts R.
Low-dose cyclosporine versus placebo in patients with
rheumatoid arthritis. Lancet 1990; 335: 1051-1055.
69. OBoyle CA, McGee H, Hickey A, 0 Malley K, Joyce
CRB. Individual quality of life in patients undergoing
hip replacement. Lancet 1992; 339: 1088-1091.
70. Meenan RF, Gertman PM, Mason JH, Dunaif R.
Measuring health status in arthritis: The Arthritis
Impact Measurement Scales. Arthritis Rheum 1980; 23:
146-152.
71. Fries JF, Spitz PW, Young DY. Measurement of
health outcome in arthritis. Arthritis Rheum 1980; 23:
137-145.
12. Flanagan JD. A research approach to improving our
quality of life. Am F%ychol 1978; 33: 138-147.

(Appendix overleaf)

FRANCIS
GUILLEMIN
et al.

1432

APPENDIX
ARSTRACIION FORM
FOR THE APPRAISAL OF CROSS-CULTURAL ADAPTATION OF HRQOL INSTRUMENTS
Originalinstrument:
Population addressed by the target version:
Language:
culture:
Country of origin:
Country of residence:
SCORE
1. Tradaiion

tedoliqoe:

Number of translations:
Number of translators in each translation:

Yes
Yes
Yes

No
No
No

Were they translating into their mother tongue:

Yes

No

Were they aware of the concepts:

Were they aware of the target condition(s):

Yes No

Were they translating into their mother tongue:

Were they aware of the concepts:
Were they aware of the target condition(s):

2. Back-tramlatilIn technique
Number of back-translations:
Number of translators in each back-translation:

Yes

No

3. committee approach

Yes No

Committee review done:

Composition of the committee:

Yes

No

With bilinguals:

Yes

No

With monolinguals:
using a probe technique:
using another technique:

Yes
Yes

No

Yes

No

Translation/back-translation

process iterated:

??

4. Pre-teat@

No

specify:
Sample composition:

??

Sample size:
5. woighthg

SCORE

seorea adoptatioo

Was the weighting of scores examincd:

with patients:
with other experts:
What method was used:
Which were the results:

Yes
Yes
Yes

No

No
No

Not applicable

??